Olympus-OM
[Top] [All Lists]

[OM] More comparisons [was Comparison shots]

Subject: [OM] More comparisons [was Comparison shots]
From: Moose <olymoose@xxxxxxxxx>
Date: Sun, 02 Nov 2008 15:37:20 -0800
Ken Norton wrote:

> I'm agreeing with Chuck, here.  

And I'm agreeing with you [cataclysmic music as image of Earth being 
torn apart as it drops into a black hole.  :-) ] - and with Chuck.

> Group B illustrates several film "problems".
>   

It's just film, focus, lens, camera (mirror slap?, ?), steadiness, 
and/or scan problem or a combination of those factors that results in an 
image far worse than it should be to be used as an exemplar of its type.

I don't have a V700. However, before buying a Canon 9950F, I looked 
closely at the reviews of both at photo-i. I came to the conclusion that 
the V700 might be capable of just a shade more detail resolution, but 
that the difference was so small as to mean nothing in practical use. My 
sense was that dynamic range was essentially the same. I picked the 
Canon based on other factors.

I've scanned Portra 400NC and it's way better than this image would 
indicate. I just scanned some on the 9950F @ 2400 dpi, using VueScan. 
This is the original 400NC, not the new stuff that I understand scans 
even better. 35mm film scanned using an ICC color profile I made with 
VueScan and an IT8 target. They were then downsampled using FM SI Pro to 
the same effective dpi as Wayne's group B image. 

As I and others have explained before, all digitally captured images, 
including film scans, require sharpening to make all the detail they 
capture visible. This is even more the case with flatbed scanners than 
dedicated film scanners. So these images have been sharpened in two 
steps, once as part of NR using NeatImage and once using FM 
IntelliSharpen II. I also made a small Curves adjustment.

I don't have any 400NC images of crowds of people, but I do have them of 
crowds of foliage and flowers, 
<http://galleries.moosemystic.net/MooseFoto/index.php?gallery=Miscellaneous&image=691_19_r1016dpi_ii.jpg>
and. 
<http://galleries.moosemystic.net/MooseFoto/index.php?gallery=Miscellaneous&image=691_17_r1016dpi_ii.jpg>

I don't know for sure, but I'd guess these would compare pretty 
favorably with images taken at the same time with an E-510. Proper 
comparison images from 6x7 400NC should beat them both for detail. I 
also scanned them @ 4800 dpi. Downsized to the same size and similarly 
processed, one 4800dpi scan had a little more detail and the other about 
the same as the lower rez scans.

> Number one issue that I see is a narrow Dmax scan which forced noise in the 
> shadows to increase too greatly. It looks like the scanner was underexposing 
> and Vuescan's curves adjustment (most likely auto) was correcting things so 
> much you didn't see where the actual scan exposure was actually at.  We must 
> remember that a digital scanner is a digital camera with the same 
> requirements and limitations in exposure that any digital camera has.
>   

It looks to me like the Black Point may not have been at 0%. I've found 
that scanning color neg. film on the 9950F, setting Black Point above 0% 
does troublesome things to deep shadow. It also appears White Point may 
have been off, leading to blocked highlights. With Vuescan and CN film, 
I've found that using Neutral Color Balance setting, BP of 0% and WP of 
0-0.1% generally gives a nice distribution of tonal range across the 
histogram. Sometimes, specular or other really out of range highlights 
will require either WP up to almost 1% and/or adjustment of the 
Brightness setting to put mid tones where they belong.

> The second thing I noticed was an incorrect adjustment for the film base. 
> When scanning negative color films, one must not just go for balancing the 
> mid-tones, but also balance for a different gamma across the entire
> spectrum.  Some scanner software does this better than others and I've pulled 
> much hair out (I have plenty more) attempting to get good scans from color 
> negs with Vuescan.  This is one area where the Nikon Scan software does a 
> phenominal job and saves me much grief.
>   

I'm not sure why some people seem to have so much trouble scanning CN 
film with VueScan. Perhaps in not understanding that the process and 
settings really need to be different? Maybe because the reversal 
reverses highlight and shadow issues? Base colors?

You can get good results for the film base either with the built in 
adjustments for many older Kodak films, or by sampling a piece of clear 
base within VueScan for any film.

The real solution to the larger issue of gamma and film toes and 
shoulders, from my perspective, is ICC profiles. These may be built 
using existing film base settings, where they exist in VueScan. For 
those and all other films, setting film film base to generic both when 
making the profile and when scanning film using it will correct for both 
base and curve at once.

The above images look pretty true to life to me with no color adjustment 
at all after scanning with a profile. As 160NC and 400NC are carefully 
balanced by Kodak to have the same color characteristics, I use a 
profile from 160NC for both.

> The third thing that jumped out at me is the yellow cast in Group B. This may 
> be totally related to my second observation, but it could also be due to the 
> notorious yellowing of the old Mamiya lenses.  I noticed that this was an 
> RB67 shot, but you did not reveal the vintage of the lens. (or maybe you did, 
> and I just don't know my model descriptions well enough. From experience with 
> my classic silver-nosed Zuikos, this yellowing does actually affect the reds 
> and not necessarily in a positive way.
>   

Once again, it would be possible to make separate profiles for lenses 
with different color characteristics. I know it may sound like a pain, 
but it's really quite easy and the advantages of time saved in scanning 
and post and the increased color accuracy are enormous.

> Through minimal adjustments I can get Group B to mimick Group A, but the 
> noise in Groub B is outrageous.  I come back around to a bad scan. What you 
> are seeing is NOT grain, but noise.  A 2400dpi scan of Portra 400NC is not 
> going to have enough grain to be an issue--especially in a downsized image.
>   

I agree, it's not grain. What combination of scanning software settings 
and post processing created it I don't know. Even at full pixel, 
4800dpi, the grain is modest and easily corrected, if desired.

> I also question the focus accuracy of the scanner. This reminds me of scans I 
> got with my old Coolscan II. 
I do seem to recall some discussion of such issues with height spacers and/or 
film holders on the V7xx scanners. As with their film scanners, Canons don't 
seem to have DOF issues.


> Of course, Wayne could be yanking our chain.
>   
If so, it worked.

Moose

==============================================
List usage info:     http://www.zuikoholic.com
List nannies:        olympusadmin@xxxxxxxxxx
==============================================

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz